摘要 |
PURPOSE:To designate a case at the deep layer structure of a variable in an input sentence while the input form, a natural Japanese language input, is held by providing a means to input the input sentence including the information to designate the case in the deep layer structure. CONSTITUTION:A Japanese language sentence inputted to a keyboard 1 is converted to KANA (Japanese sentence)/KANJI (Chinese character) by a Japanese language interface 3 and a reasoning module 4 operated by the control of a CPU 2, and the response based upon the Japanese language sentence is prepared and displayed at a CRT 5. In such a case, an input sentence is first converted through the morpheme processing part of the interface 3 to a surface layer structure, stored into the surface layer structure buffer, and thereafter, converted to the deep layer structure by a deep layer structure extracting part 4a. In such a case, it is judged whether or not the part of speech included in the surface layer is a variable composed of a special character column for designating the deep layer case, and to the deep layer structure, the deep layer case designated by the variable is given. Thus, a mistake on a reasoning can be prevented, and by inputting a natural input sentence, the case at the deep layer structure can be clearly designated.
|